arXiv subjects

Jing Liu

Publications and source records attributed to Jing Liu.

At least 253 records · Page 14Linked to original sources

arXiv2025

LatentLLM: Attention-Aware Joint Tensor Compression

Modern foundation models such as large language models (LLMs) and large multi-modal models (LMMs) require a massive amount of computational and memory resources. We propose a new framework to convert such LLMs/LMMs into a reduced-dimension latent structure. Our method extends a local activation-aware tensor decomposition to a global attention-aware joint tensor de-composition. Our framework can significantly improve the model accuracy over the existing model compression methods when reducing the latent dimension to realize computationally/memory-efficient LLMs/LLMs. We show the benefit on several benchmark including multi-modal reasoning tasks.

cs.LG↗

arXiv2025

AniSora: Exploring the Frontiers of Animation Video Generation in the Sora Era

Animation has gained significant interest in the recent film and TV industry. Despite the success of advanced video generation models like Sora, Kling, and CogVideoX in generating natural videos, they lack the same effectiveness in handling animation videos. Evaluating animation video generation is also a great challenge due to its unique artist styles, violating the laws of physics and exaggerated motions. In this paper, we present a comprehensive system, AniSora, designed for animation video generation, which includes a data processing pipeline, a controllable generation model, and an evaluation benchmark. Supported by the data processing pipeline with over 10M high-quality data, the generation model incorporates a spatiotemporal mask module to facilitate key animation production functions such as image-to-video generation, frame interpolation, and localized image-guided animation. We also collect an evaluation benchmark of 948 various animation videos, with specifically developed metrics for animation video generation. Our entire project is publicly available on https://github.com/bilibili/Index-anisora/tree/main.

cs.GR↗

arXiv2025

Model Merging in Pre-training of Large Language Models

Model merging has emerged as a promising technique for enhancing large language models, though its application in large-scale pre-training remains relatively unexplored. In this paper, we present a comprehensive investigation of model merging techniques during the pre-training process. Through extensive experiments with both dense and Mixture-of-Experts (MoE) architectures ranging from millions to over 100 billion parameters, we demonstrate that merging checkpoints trained with constant learning rates not only achieves significant performance improvements but also enables accurate prediction of annealing behavior. These improvements lead to both more efficient model development and significantly lower training costs. Our detailed ablation studies on merging strategies and hyperparameters provide new insights into the underlying mechanisms while uncovering novel applications. Through comprehensive experimental analysis, we offer the open-source community practical pre-training guidelines for effective model merging.

cs.CL↗

arXiv2025

Emergent Specialization: Rare Token Neurons in Language Models

Large language models struggle with representing and generating rare tokens despite their importance in specialized domains. In this study, we identify neuron structures with exceptionally strong influence on language model's prediction of rare tokens, termed as rare token neurons, and investigate the mechanism for their emergence and behavior. These neurons exhibit a characteristic three-phase organization (plateau, power-law, and rapid decay) that emerges dynamically during training, evolving from a homogeneous initial state to a functionally differentiated architecture. In the activation space, rare token neurons form a coordinated subnetwork that selectively co-activates while avoiding co-activation with other neurons. This functional specialization potentially correlates with the development of heavy-tailed weight distributions, suggesting a statistical mechanical basis for emergent specialization.

cs.AI↗

arXiv2025

Enhancement of small-scale induced gravitational waves from the soliton/oscillon domination

We investigate density fluctuations and scalar-induced gravitational waves (GWs) arising from the production of long-lived solitons and oscillons, which can dominate the early Universe and drive reheating prior to the standard radiation-dominated era. Curvature perturbations are generated not only by the Poisson distribution of these solitons/oscillons but are also amplified during the early matter-dominated phase when the solitons/oscillons dominate. Scalar-induced GWs with characteristic energy spectra emerge from these amplified curvature perturbations, particularly at the sudden transition from matter domination to radiation domination. We analyze this scenario and its GW predictions in detail, focusing on the impact of the initial energy fraction, the lifetime, and the average separation of these solitons/oscillons. This provides a wealth of potential observational targets for various GW detectors. Furthermore, for inflation and preheating models that produce oscillons with large separations, we derive new constraints on these models based on the upper bounds on the effective number of relativistic degrees of freedom, as anticipated from future cosmic microwave background and large-scale structure observations.

astro-ph.CO↗

arXiv2025

Scaling Law for Quantization-Aware Training

Large language models (LLMs) demand substantial computational and memory resources, creating deployment challenges. Quantization-aware training (QAT) addresses these challenges by reducing model precision while maintaining performance. However, the scaling behavior of QAT, especially at 4-bit precision (W4A4), is not well understood. Existing QAT scaling laws often ignore key factors such as the number of training tokens and quantization granularity, which limits their applicability. This paper proposes a unified scaling law for QAT that models quantization error as a function of model size, training data volume, and quantization group size. Through 268 QAT experiments, we show that quantization error decreases as model size increases, but rises with more training tokens and coarser quantization granularity. To identify the sources of W4A4 quantization error, we decompose it into weight and activation components. Both components follow the overall trend of W4A4 quantization error, but with different sensitivities. Specifically, weight quantization error increases more rapidly with more training tokens. Further analysis shows that the activation quantization error in the FC2 layer, caused by outliers, is the primary bottleneck of W4A4 QAT quantization error. By applying mixed-precision quantization to address this bottleneck, we demonstrate that weight and activation quantization errors can converge to similar levels. Additionally, with more training data, weight quantization error eventually exceeds activation quantization error, suggesting that reducing weight quantization error is also important in such scenarios. These findings offer key insights for improving QAT research and development.

cs.LG↗

arXiv2025

From Weak Labels to Strong Results: Utilizing 5,000 Hours of Noisy Classroom Transcripts with Minimal Accurate Data

Recent progress in speech recognition has relied on models trained on vast amounts of labeled data. However, classroom Automatic Speech Recognition (ASR) faces the real-world challenge of abundant weak transcripts paired with only a small amount of accurate, gold-standard data. In such low-resource settings, high transcription costs make re-transcription impractical. To address this, we ask: what is the best approach when abundant inexpensive weak transcripts coexist with limited gold-standard data, as is the case for classroom speech data? We propose Weakly Supervised Pretraining (WSP), a two-step process where models are first pretrained on weak transcripts in a supervised manner, and then fine-tuned on accurate data. Our results, based on both synthetic and real weak transcripts, show that WSP outperforms alternative methods, establishing it as an effective training methodology for low-resource ASR in real-world scenarios.

eess.AS↗

arXiv2025

Unveiling Knowledge Utilization Mechanisms in LLM-based Retrieval-Augmented Generation

Considering the inherent limitations of parametric knowledge in large language models (LLMs), retrieval-augmented generation (RAG) is widely employed to expand their knowledge scope. Since RAG has shown promise in knowledge-intensive tasks like open-domain question answering, its broader application to complex tasks and intelligent assistants has further advanced its utility. Despite this progress, the underlying knowledge utilization mechanisms of LLM-based RAG remain underexplored. In this paper, we present a systematic investigation of the intrinsic mechanisms by which LLMs integrate internal (parametric) and external (retrieved) knowledge in RAG scenarios. Specially, we employ knowledge stream analysis at the macroscopic level, and investigate the function of individual modules at the microscopic level. Drawing on knowledge streaming analyses, we decompose the knowledge utilization process into four distinct stages within LLM layers: knowledge refinement, knowledge elicitation, knowledge expression, and knowledge contestation. We further demonstrate that the relevance of passages guides the streaming of knowledge through these stages. At the module level, we introduce a new method, knowledge activation probability entropy (KAPE) for neuron identification associated with either internal or external knowledge. By selectively deactivating these neurons, we achieve targeted shifts in the LLM's reliance on one knowledge source over the other. Moreover, we discern complementary roles for multi-head attention and multi-layer perceptron layers during knowledge formation. These insights offer a foundation for improving interpretability and reliability in retrieval-augmented LLMs, paving the way for more robust and transparent generative solutions in knowledge-intensive domains.

cs.CL↗

arXiv2025

End-to-End Vision Tokenizer Tuning

Existing vision tokenization isolates the optimization of vision tokenizers from downstream training, implicitly assuming the visual tokens can generalize well across various tasks, e.g., image generation and visual question answering. The vision tokenizer optimized for low-level reconstruction is agnostic to downstream tasks requiring varied representations and semantics. This decoupled paradigm introduces a critical misalignment: The loss of the vision tokenization can be the representation bottleneck for target tasks. For example, errors in tokenizing text in a given image lead to poor results when recognizing or generating them. To address this, we propose ETT, an end-to-end vision tokenizer tuning approach that enables joint optimization between vision tokenization and target autoregressive tasks. Unlike prior autoregressive models that use only discrete indices from a frozen vision tokenizer, ETT leverages the visual embeddings of the tokenizer codebook, and optimizes the vision tokenizers end-to-end with both reconstruction and caption objectives. ETT can be seamlessly integrated into existing training pipelines with minimal architecture modifications. Our ETT is simple to implement and integrate, without the need to adjust the original codebooks or architectures of the employed large language models. Extensive experiments demonstrate that our proposed end-to-end vision tokenizer tuning unlocks significant performance gains, i.e., 2-6% for multimodal understanding and visual generation tasks compared to frozen tokenizer baselines, while preserving the original reconstruction capability. We hope this very simple and strong method can empower multimodal foundation models besides image generation and understanding.

cs.CV↗

arXiv2025

On the Account Security Risks Posed by Password Strength Meters

Password strength meters (PSMs) have been widely used by websites to gauge password strength, encouraging users to create stronger passwords. Popular data-driven PSMs, e.g., based on Markov, Probabilistic Context-free Grammar (PCFG) and neural networks, alarm strength based on a model learned from real passwords. Despite their proven effectiveness, the secure utility that arises from the leakage of trained passwords remains largely overlooked. To address this gap, we analyze 11 PSMs and find that 5 data-driven meters are vulnerable to membership inference attacks that expose their trained passwords, and seriously, 3 rule-based meters openly disclose their blocked passwords. We specifically design a PSM privacy leakage evaluation approach, and uncover that a series of general data-driven meters are vulnerable to leaking between 10^4 to 10^5 trained passwords, with the PCFG-based models being more vulnerable than other counterparts; furthermore, we aid in deriving insights that the inherent utility-privacy tradeoff is not as severe as previously thought. To further exploit the risks, we develop novel meter-aware attacks when a clever attacker can filter the used passwords during compromising accounts on websites using the meter, and experimentally show that attackers targeting websites that deployed the popular Zxcvbn meter can compromise an additional 5.84% user accounts within 10 attempts, demonstrating the urgent need for privacy-preserving PSMs that protect the confidentiality of the meter's used passwords. Finally, we sketch some counter-measures to mitigate these threats.

cs.CR↗

arXiv2025

Optimal time-decay for Euler-Fourier system with damping in the critical $L^2$ framework

This paper is concerned with the large time behavior of solutions to the Euler-Fourier system with damping in $\mathbb{R}^{d}~(d\geq1)$. A time-weighted energy argument has been developed within the $L^2$ framework to derive the optimal time-decay rates, which enables us to remove the smallness of low-frequencies of initial data. A great part of our analysis relies on the study of a Lyapunov functional in the spirit of [13], which mainly depends on some elaborate use of non-classical Besov product estimates and interpolations. Exhibiting a damped mode with faster time decay than the whole solution also plays a key role.

math.AP↗

arXiv2025

A combined analysis of $Y(2240)$ in the $e^+ e^- \to K^+ K^-$ and $e^+ e^- \to K_S^0 K_L^0$ processes

In the present work, we perform a combined analysis of the cross sections for $e^+ e^- \to K^+ K^-$ and $e^+ e^- \to K_S^0 K_L^0$ by including the interference between direct coupling process and resonance intermediate process. When setting the isospin factor as a free parameter, we can well reproduce the structures in the cross sections for $e^+ e^- \to K^+ K^-$ and $e^+ e^- \to K_S^0 K_L^0$, however, the isospin factor deviates with the isospin symmetry conservation predicted value of one. By fixing the isospin factor to be one, we find the cross sections for $e^+ e^-\to K^+ K^-$ can be accurately fitted, while the same model parameters fail to adequately reproduce the observed cross section for the process $e^+ e^- \to K_S^0 K_L^0$, which indicate that there should be contributions from other vector meson states, such as $ρ(2150)$ and $ω(2220)$. More precise experimental measurement below 2.0 GeV and between $2.3\sim 2.5$ GeV may provide more information about the vector resonances in this energy range, which should be accessible for the BESIII collaboration.

hep-ph↗

arXiv2025

Top Score on the Wrong Exam: On Benchmarking in Machine Learning for Vulnerability Detection

According to our survey of machine learning for vulnerability detection (ML4VD), 9 in every 10 papers published in the past five years define ML4VD as a function-level binary classification problem: Given a function, does it contain a security flaw? From our experience as security researchers, faced with deciding whether a given function makes the program vulnerable to attacks, we would often first want to understand the context in which this function is called. In this paper, we study how often this decision can really be made without further context and study both vulnerable and non-vulnerable functions in the most popular ML4VD datasets. We call a function "vulnerable" if it was involved in a patch of an actual security flaw and confirmed to cause the program's vulnerability. It is "non-vulnerable" otherwise. We find that in almost all cases this decision cannot be made without further context. Vulnerable functions are often vulnerable only because a corresponding vulnerability-inducing calling context exists while non-vulnerable functions would often be vulnerable if a corresponding context existed. But why do ML4VD techniques achieve high scores even though there is demonstrably not enough information in these samples? Spurious correlations: We find that high scores can be achieved even when only word counts are available. This shows that these datasets can be exploited to achieve high scores without actually detecting any security vulnerabilities. We conclude that the prevailing problem statement of ML4VD is ill-defined and call into question the internal validity of this growing body of work. Constructively, we call for more effective benchmarking methodologies to evaluate the true capabilities of ML4VD, propose alternative problem statements, and examine broader implications for the evaluation of machine learning and programming analysis research.

cs.CR↗

arXiv2025

Quantum-enhanced weak absorption estimation with correlated photons

Conventional absorption spectroscopy relies on coherent laser sources, and in turn suffers from the inherent limitation of shot noise, especially in estimating weak absorption. Here we propose a measurement strategy with correlated photons to determine the weak absorption by distinguishing the output with and without photons, dubbed as the on-off measurement. We demonstrate that absorption spectroscopy that incorporates quantum correlations is capable of estimating weak absorption down to a single-photon level, even in noisy environments, achieving a precision comparable to that obtained through 1000 photons in conventional absorption spectroscopy. Our strategy provides a new method to probe fragile systems with weak absorption, avoiding the occurrence of light-induced damage.

quant-ph↗

arXiv2025

A Unified Framework for Iris Anti-Spoofing: Introducing Iris Anti-Spoofing Cross-Domain-Testing Protocol and Masked-MoE Method

Iris recognition is widely used in high-security scenarios due to its stability and distinctiveness. However, iris images captured by different devices exhibit certain and device-related consistent differences, which has a greater impact on the classification algorithm for anti-spoofing. The iris of various races would also affect the classification, causing the risk of identity theft. So it is necessary to improve the cross-domain capabilities of the iris anti-spoofing (IAS) methods to enable it more robust in facing different races and devices. However, there is no existing protocol that is comprehensively available. To address this gap, we propose an Iris Anti-Spoofing Cross-Domain-Testing (IAS-CDT) Protocol, which involves 10 datasets, belonging to 7 databases, published by 4 institutions, and collected with 6 different devices. It contains three sub-protocols hierarchically, aimed at evaluating average performance, cross-racial generalization, and cross-device generalization of IAS models. Moreover, to address the cross-device generalization challenge brought by the IAS-CDT Protocol, we employ multiple model parameter sets to learn from the multiple sub-datasets. Specifically, we utilize the Mixture of Experts (MoE) to fit complex data distributions using multiple sub-neural networks. To further enhance the generalization capabilities, we propose a novel method Masked-MoE (MMoE), which randomly masks a portion of tokens for some experts and requires their outputs to be similar to the unmasked experts, which can effectively mitigate the overfitting issue of MoE. For the evaluation, we selected ResNet50, VIT-B/16, CLIP, and FLIP as representative models and benchmarked them under the proposed IAS-CDT Protocol.

cs.CV↗

arXiv2025

QuanEstimation.jl: An open-source Julia framework for quantum parameter estimation

As the main theoretical support of quantum metrology, quantum parameter estimation must follow the steps of quantum metrology towards the applied science and industry. Hence, optimal scheme design will soon be a crucial and core task for quantum parameter estimation. To efficiently accomplish this task, software packages aimed at computer-aided design are in high demand. In response to this need, we hereby introduce QuanEstimation.jl, an open-source Julia framework for scheme evaluation and design in quantum parameter estimation. It can be used either as an independent package or as the computational core of the recently developed hybrid-language (Python-Julia) package QuanEstimation [Phys. Rev. Res. 4 (4) (2022) 043057]. Utilizing this framework, the scheme evaluation and design in quantum parameter estimation can be readily performed, especially when quantum noises exist.

quant-ph↗

arXiv2025

Networking Systems for Video Anomaly Detection: A Tutorial and Survey

The increasing utilization of surveillance cameras in smart cities, coupled with the surge of online video applications, has heightened concerns regarding public security and privacy protection, which propelled automated Video Anomaly Detection (VAD) into a fundamental research task within the Artificial Intelligence (AI) community. With the advancements in deep learning and edge computing, VAD has made significant progress and advances synergized with emerging applications in smart cities and video internet, which has moved beyond the conventional research scope of algorithm engineering to deployable Networking Systems for VAD (NSVAD), a practical hotspot for intersection exploration in the AI, IoVT, and computing fields. In this article, we delineate the foundational assumptions, learning frameworks, and applicable scenarios of various deep learning-driven VAD routes, offering an exhaustive tutorial for novices in NSVAD. In addition, this article elucidates core concepts by reviewing recent advances and typical solutions and aggregating available research resources accessible at https://github.com/fdjingliu/NSVAD. Lastly, this article projects future development trends and discusses how the integration of AI and computing technologies can address existing research challenges and promote open opportunities, serving as an insightful guide for prospective researchers and engineers.

cs.CV↗

arXiv2025

Image Difference Grounding with Natural Language

Visual grounding (VG) typically focuses on locating regions of interest within an image using natural language, and most existing VG methods are limited to single-image interpretations. This limits their applicability in real-world scenarios like automatic surveillance, where detecting subtle but meaningful visual differences across multiple images is crucial. Besides, previous work on image difference understanding (IDU) has either focused on detecting all change regions without cross-modal text guidance, or on providing coarse-grained descriptions of differences. Therefore, to push towards finer-grained vision-language perception, we propose Image Difference Grounding (IDG), a task designed to precisely localize visual differences based on user instructions. We introduce DiffGround, a large-scale and high-quality dataset for IDG, containing image pairs with diverse visual variations along with instructions querying fine-grained differences. Besides, we present a baseline model for IDG, DiffTracker, which effectively integrates feature differential enhancement and common suppression to precisely locate differences. Experiments on the DiffGround dataset highlight the importance of our IDG dataset in enabling finer-grained IDU. To foster future research, both DiffGround data and DiffTracker model will be publicly released.

cs.CV↗